The FORMULATION of PRUSSIAN BLUE

PRUSSIAN BLUE. DAVID DAVIDSON. Bmoklyn College, Brooklyn, New York. The several blue precipitates obtained from simple iron and ferrocyanide ions, the...
1 downloads 0 Views 4MB Size
of

The FORMULATION

PRUSSIAN BLUE DAVID DAVIDSON Bmoklyn College, Brooklyn, New York

The several blue precipitates obtained from simple iron ions and the complex iron cyanides may be formulated as deriuatiwes of a $olynuclear complex composed of ferric

and ferrocyanide ions, the formation of which involves co6rdination by both ends of the cyanide ion.

++++++

I

N a previous paper i t was shown that equimolecular quantities of ferric and ferrocyanide ions, or of ferrous and ferricyanide ions, react instantaneously to form an equilibrium mixture of all four ions according to the equation Fe+++ [ F e ( C N ) B ] ' z F e + + [Fe(CN)6]'; and that the formation of a blue product is subsequent to the establishment of this equilibrium.' It is the object of the present paper to answer such questions as, what ions are involved in the formation of the blue product? How does the composition of Prussian blue vary when excess of ferric or of ferrous ions are employed? To what may the common blne color of these products be ascribed?

and his coworkers concluded that it was a ferrocyanide because it was produced from ferric ferricyanide by the action of hydrogen peroxide which is known to reduce ferricyanide, but not ferric ion. On the other hand, Eibner and Gerstacker4 considered Prussian blue a ferricyanide since it was formed from femc ferricyanide by the action of sulfur dioxide which reduces ferric ion, but not ferricyanide. Like the arguments based on decomposition, these results are invalidated by the existence of the redox equilibrium.6 It is now possible to offersupport for the ferrocyanide nature of soluble Prnssian blue, which, since it is based upon the redox equilibrium, avoids the objection mentioned above. Since the formation of soluble SOLUBLE PRUSSIAN BLUE Prnssian blue is relatively slow, some light may be Analysis of the blue product obtained from equi- thrown on the question of its composition by noting molecular quantities of ferric and ferrocyanide ions, the effect of excess of each of the ions involved in the or from f m o u s and femcyanide ions, indicates that redox equilibrium upon the time required for blue it contains one alkali metal ion, one iron ion, and one formation. As has been demonstrated previously,l complex hexacyanide ion; that i t is either KFe+++- excess of ferrocyanide, femcyanide, or of ferrous ions [Fe(CN)#lSor KFe++[Fe(CN)fi]-. Several methods each accelerate blue formation, whiie excess of ferric Prussian blue paradox. have been employed to distinguish between these ion actually retards it-the formulas. The first which comes to mind is decompo- While the interpretation of the accelerating effect of sition. Thus sodium hydroxide converts soluble the first-mentioned ions is not free from ambiguity Prussian blue to femc hydroxide and sodium ferro- the inhibiting effect of ferric ion is illuminating, for cyanide. Such evidence of the ferrocyanide nature the only ionic factor which is decreased by excess of of the iron blue is invalid, however, because of the ferric ion is ferrocyanide ion. This inverse correlation operation of the mobile redox equilibrium given above. between excess of ferric ion and the rate of formation According to the solubility or complexity of a de- of Prussian blue leads to the unavoidable conclusion composition product containing one or the other ions, that Prussian blne is derived from ferrocyanide ion. Other less direct forms of argument may be applied it should be possible to drive this oxidation-reduction reaction to completion in either direction. How this to this problem. Ferric salts are generally found to actually occurs and leads to conflicting testimony is be less soluble than the compondimg ferrous salts. illustrated by the claim of Woringerx that ferricyanide Similarly, ferrocyanides are generally less soluble than ion is common to the Prnssian blues, since their de- the corresponding femcyanides. Hence, of the four composition with ammonium carbonate yields am- possible ionic combinations (ferrous ferrocyanide, femc ferricyanide, ferrous femcyanide, and ferric monium femcyanide. Other writers have attempted to amve a t the com- ferrocyanide), ferrous ferricyanide should be expected position of soluble Prussian blue by considering the to be the most soluble salt and ferric ferrocyanide the methods by which it may be produced. HofmannS least soluble. Consequently, Prnssian blue, which is practically insoluble in water (the term "soluble F'rnssian blue" is a misnomer connoting the ease with

+

+

-

'

A. E ~ N EAND R L. GERSTACKER, Chem.-Ztg., 3'1,137,178,195 (1913). E. MWLER, ibid., 38, 281,328 (1914).

'7

which this substance forms colloidal solutions), must be femc ferrocyanide rather than ferrous ferricyanide. If oxidation-reduction did not intervene, ferrous femcyanide would be more soluble than ferric femcyanide which is so excessively soluble in water as to be practically unknown in the solid state.6 Further evidence in favor of considering soluble Prussian hlue a ferrocyanide may be derived from Kolthoff's work' on the precipitation of heavy metals by ferrocyanide and by femcyanide. Whereas the ferrocyanides of silver, copper, and zinc, for example, showed a marked tendency to form double salts with potassium ferrocyanide, the corresponding ferricyanides were obtained only as simple salts. Furthermore, the colors of these ferricyanides all corresponded to the colors of the component ions in contrast to the marked difference between the color of Prussian hlue and the ions giving rise to it. PRUSSIAN BLUE AS A POLYNUCLEAR COMPLEX

The foregoing discussion makes i t reasonable to consider soluble Prussian hlue as potassium ferric ferrocyanide, KFe+++[Fe(CN)6]-. But how does i t happen that such a deep hlue salt is formed from practically colorless ions? The answer-omplex formation-readily suggests itself. Unfortunately, reliable methods of studying complexes (such as conductivity, solubility, and potentiometry) are limited to soluble sub~tances.~X-ray analysis may be of some value in studying the complexity of solids, but its failure in the simpler case of silver cyanide vs. silver argenticyanide makes its utility here dubious.9 With iron salts, magnetic susceptibilities may be indicative of complex formation. Davidson and Welo,'a however, were unable to demonstrate the existence of a polynuclear complex in Prussian blue by this method. Their failure should not he iuterpreted as evidence against such a structure since, for example, the behavior of (FeF6)' which shows a Weiss magneton number of 29 is inconsistent with its formal similarity to [Fe(CN)6]' which exhibits a magneton number of 12." Lacking more reliable methods of confirming the complex indicated by the color of Prussian blue, recourse will be had in the sequel to an argument based on the composition of the several iron cyanide blues. This will be prefaced by a theoretical analysis of the possibilities for supercomplex formation from ferrocyanide ion.

-

H. REIRLEN AND U. V.KUMMER, Ann., 469,43 (1929) claim, however, to have prepared an insoluble isomer of ferric ferri-

-. I. M. KOLTUOFF, 2.

,.~..,irl~~ -> 7

The electronic structure of cyanide ion may be written as :C:::N:, :C::N::, or : : C : N : : . ' V n all of these formulas a t least one lone pair of electrons is present on both the carbon and nitrogen atoms. In the formation of ferrocyanide from ferrous ion and six cyanide groups, i t appears to be the lone pair of the carbon atom which is involved in the coordination bond. This is supported by the fact that alkylation of ferrocyanide yields an isonitrile complex, whereas the alkylation of uncoordinated cyanides (sodium cyanide) yields mainly nitrile. I t is conceivable, therefore, that a lone pair remaining on the nitrogen atoms of these cyanide groups may function to coordinate metals; in other words, that a ferrocyanide ion may satisfy siz coordination positions of simple metallic ions. It should he borne in mind that ferrocyanide ion is an exceedingly stable structure which shows practically no tendency to dissociate into simpler ions. Hence, ferric ion with a coordination number of six may be coordinated by a ferrocyanide ion thus [[Fe(CN)6]FFe]. It is evident that spatial limitations prevent one ferric ion from being coordinated with one ferrocyanide ion by six bonds. Current stereochemical conceptions of complex salts would limit the number of such bonds to three. Structures should, therefore, be considered in which a ferric ion is coordinated with one ferrocyanide ion by one, two, or three bonds, the six bonds being obtained by involving six, three, or two ferrocyanide ions, respectively. These cases will be discussed in turn. (a) Three-bond coordination (fused 6-rings).-If three of the cyanide groups of one ferrocyanide ion serve to occupy three coordination positions of a single femc ion there results the incomplete structure

This unit, containing a system of fused six-membered rings, has three unsatisfied coordination positions on the ferric ion and three uncoordinated nitrogen atoms. While steric limitations prevent the completion of coordination within this unit, inter-unit coordination may occur by a head-to-tail union of these groups. This process resembles the polymerization of unsaturated organic compounds, such as vinyl chloride, for e~ample.'~Thus, a linear polymer, :EF~++/E!~F~+++~~$~F~++LE~~F~+++L~~. NC/

\CN/ -. .

\NC/

\CN/ -. .

\

a d . Chem., 62, 209 (1923).

0 WM. THOMAS, "Complex salts," Blackie and Son Ltd., London, 1924, Chap. 11. L. P. HAMMETT, "Solutions of electrol~tes,"2nd edition, McGraw-Hill Book Company, New York City, 1934 P. 120. A nrelrmmarv reoort of the X-rav analvsis of Prussian hlue ha; recently &&red [J. F. KEGG;N AND-F. D. MILES. Nature, 137,577 (1936)l. 10 D. DAVIDSON AND L. A. WEW, I. Pkys. Chem., 32, 1191 (1928). 0.BAUDISCH AND L. A. WELO, Chcm. R&ws, 15,11(1934).

may be formed in which successive iron atoms are distinguished not only by their charges but also by the type of coordination bond (c-F~ or N-F~), --

'% W. M. LATIMER AND J. H. HILDEBRAND, "Reference book of inorganic chemistry." The Macmillan Company, New York City, 1929, p. 227. l 3 W. H. CAROTHERS, Chem. Rm'ms, 8,353 (1931).

(b) Two-bond coordination (&membered rings).Should only two of the cyanide groups of a single ferrocyanide ion be considered to be coiirdinated by a single ferric ion there would be obtained the unit

which contains a six-membered ring and four vacant positions on the ferric ion as well as four uncoordinated nitrogens. Here, too, complete coiirdmation requires polymerization although in a more complicated fashion in three dimensions only inadequately represented as foIlows:

structure similar to that given in (a),,whiie Reihlen and Z i m m e ~ m a nhave ~ ~ suggested dismembering the ferrocyanide ion to form

These authors, however, did not adhere to the following principles which, it is submitted, are involved in the formation of such polynuclear complexes from ferrocyanide. 1. The ferrocyanide complex is stable and generally remains intact. 2. Its cyanide groups may coadinate at both ends. Since the carbons of the cyanide groups are attached to the central iron atom, the six nitrogen atoms remain to be coordinated by other metallic atoms. 3. The N-metal bond is relatively weak and is easily broken, for example, by alkali. 4. One, two, or three cyanide groups in a ferrocyanide ion may coordinate a single second metallic ion. THE BERLINATES

(c) One-bond co6rdination (12-membered ring).If a single ferrocyanide ion contributes no more than one lone pair to any one iron atom, the unit becomes

Polymerization of this unit could give rise to a network of twelve-membered rings (really squares), in which the ferrous and ferric ions would occupy alternate points in a cubic lattice. A cross-section of such an arrangement is given. Only four bonds are shown in this figure, the remaining two being before and behind the cross-section and perpendicular to it. N C

1 I

C N

I

N C

I

NC-FettCN-Fet++-NC-Fet+-CN I

I

f4

N

C

C

N

N C

C

N

N C

C

I 1 CN-Fe+++-NC-Fet+-CN-Feti+-NC I I I

I

etc.

Of these representations, (c) is consistent with the preliminary results of X-ray analysis published by Keggin and Miles,= and with the straight-line formulation of cyanide ion, :C:::N:, while (b) appeals because of its six-membered rings, although these would require the angular bivalent C formula for cyanide ion, :C::N::. For ordinary purposes, however, the non-committal formula, [[Fe++(CN)61EFe+++], is a sufficient and satisfactory representation of the unit resent in the ~olvnuclearcomplex of Prnssian blue. ~ a m b iand cferici14 have suggested a polynuclear . ,

-

" L. Cams=m o A. CLERICI, Gazz. chim. ilal., 58, 61 (1928).

Soluble Prnssian blue may now be considered the potassium salt of a complex acid, H+[[Fe(CN)6]%-, or fem-ferrocyanic acid. For convenience, the name "berlinic acid" is suggested for this substance. This is derived from the German name for Prussian blue, which is Berlinerblau. Accordindv. soluble Prussian blue would be potassium berlinat; ' When ferrocyanide is treated with excess of ferric ion, each ferrocyanide ion coordinates one ferric ion to form the univalent berlinate iron unit which is then precipitated by the excess of ferric ion as ferric berlinate, Fe+++[[Fe(CN)6]%e]3-, commonly written F Q + + + [ F ~ ( C N ) &(insoluble ~ Prussian blue). If excess of ferrous ion is employed on ferricyanide ion, then each ferricyanide ion reacts with one ferrous ion to form ferric and ferrocyanide ions which combine to form the berlinate ion which is then precipitated by the excess of ferrous ion as ferrous berliuate, Fe++[[Fe(CN)s]SFe]~-, or ferrous ferri-ferrocyanide, rather than the Fe3++[Fe(CN)s]z', ferrous ferricyauide, commonly and erroneously cited in textbooks (Turnbull's blue). As ordinarily precipitated Turnbull's blue is a double potassium ferrous berlmate varying from KFe(B)3 to KFez(B)5.16 TON DE M- . n -N s T. R. A.T En .--. ... . . .. .-

RRRTJNATE --

The representation of insoluble Prussian blue as ferric berlinate, Fe+++[[Fe(CN)6]zFe]a,indicates that of the four femc ions involved in its formula, three are coiirdinated by the three ferrocyanide ions, while the fourth is a simple ferric ion. This relationship may be demonstrated by the experiment which is described. For the sake of analogy the less com-

,

R S ~ L E NAND W. ZQIMERYAN. Ann.. 451, 75 (1926); 475,101 (1929). ' 6 E . MULLER, 3. prakl. Chem., (2).84,241 (1911).

280

JOURNAL OF

plicated case of silver argenticyanide is performed Silacr Argmlicyani& To 25 ce. of 0.01 M A=+. Add 10 N.. of 0.02 M

Prussinn Blur To25cc. ofO.O1 M F e t + +(in tenthformal KHSO*). Add 15ce.of 0 01 M iFe(CN)aI' (in

(CN)..

u.nr

A blue precipitate forms. This may be explained by: Feiii [Fe(CN)aI" -- (FSlFe

+

CHEM~CAL EDUCATION

colored precipitates, the one with the ferrocvanide apparently coordinated by ferric ion (berlinate ion), the other with the ferrocyanide coordinated by cupric ion (cupriferrocyanide ion). Whichever ion enters the supercomplex f i s t appears to remain intrenched, for these suspensions may be kept for years without any noticeable change in appearance.

(CN)all-, I[Fc(CN)dSFala-.

OTHER COMPLEX FERROCYANIDES

F e + + ++3[[Fe(CN)11sel=Fe++*

The principles which have been stated above goveming the formation of a polynuclear complex from ferric and ferrocyanide ions may be applied to the formulation of other complex metallic ferrocyanides. Thus, zinc ferrocyanide, Znz[Fe(CN)6]z, reacts with mart*. E;[F~(CN)G]to form KeZns[Fe(CN)~]z(Kolthoff7). (1) Comol. (2) Add 10 cc. of 0.01 M ferroThis formula becomes understandable when it is cyanide; Blfer each part. noted that ferrocvanide has a valence of f o u but (1) yields a bluc residue and a io~oriesrfiltrate; supplies six nitrogen atoms for coordination, while (2) yields no msidtte and a b1,luc filtrate. zinc ion has a valence of two and requires four groups I n (21: Fe+++[lF4CNlilEFcIr + = 2 lAg(CN)al+ IFe(CN)81' = 4[[Fe(CNl~lE for coordination. Hence, the simplest ratio of zinc Fel to ferrocyanide which will satisfy the coordination requirement is 3Zn:2[Fe(CN)6], since two ferrocyanide Volzhin" showed by quantitative tests that as long ions supply twelve nitrogens and three zinc ions have as the ratio of ferric ion to ferrocyanide was greater an aggregate coordination number of twelve. The than one, a precipitate was present, but that this unit of the polymerized polynuclear zincoferrocyanide "dissolved" as soon as the ratio became equal to unity. In 0.06 M solution, the end-point could be determined complex may, therefore, be written within 0.2 per cent. Although soluble Prussian blue does not form a true solution, nevertheless, there is a definite stoichiometric relationship between the ferric ion contained in insoluble Prussian blue and the ferrocyanide required to peptize it.

The predpitate mntaino both simple silver ion and mmplcx argentieganide ion. The uneosrdinated silver ion may be moverted t o argenticyanide ion b" addine .e x c e s of cyanide. Divide the suspension into two equal Pacts: (1) Control. (2) Add 10 ce. of 0.02 M cuanide: filier each part. (1) yields a white residue and a cyanide-free filtrate; (2) yields no residue and a n argeotieyaoidbcontaining filtrate. I n (21: Ag+[Ap(CN)zI2(CN)-

The prmipitate contains both simple ferric ion and complex ferri-lerrocyanide (berlinate) ion. The uncohdinated ferrie ion may be eonverted t o berlioate ion by adding excess of krroeyanide Divide thenuspension into two equal

FIRST COME-FIRST

SERVED

~ ~ demonstration ~ t which h suggests ~ the ~ coordinated nature of Prussian blue follows. Prursinn BIuc To 25 m. of 0.01 M [Fe(CN)al". ~ d 25 d ee.of 0.01M F ~ + +(peee * alum in tenth-formal KHS03. This produces a blue sol of soluble ~~~~~i~~blue. Add 25 cc. of 0.02 M Cut+. This produces a blue precipitate.

Knfchdl's B r m n To 25 N..of 0.01 M [Fe(CN)rl'. AM 25 OF 0.02 M cu++ (C~SO, in teoth-formal KHSOd. This produces a brown precipitate of x a t e h e t t ' ~brown. Add 25 cc. of 0.01 MFet++. This causes no apparent change.

*.

Cupric ferrocyanide forms a complex, &Cw[Fe(cN)&, which is entirely analogous to the zinc comFerrous ion with a coordination number of plex. six forms KaFe[Fe(CN)a]. Silver ferrocyanide reacts with potassium ferrocyanide to form KAg,[Fe(CN)J. This is accounted for by silver's coordination number (*g: = 6; ferrOcyanide = '.) Its polynuclear complex may differ from those prebeing P ~ ~ thus:Y ~ viOusl~discussed in

It is thus seen that two mixtures of ferrocyanide, iron, and copper ions of exactly the same composition, but prepared in different order, contain different V. VOLZSIN, J . Russ. Phys.-Ckm. Soc.. 40,480 (1908). TABLE I TABU$ or. TBB P a o s s r ~ vBLURS C m m m Name ~

~

Soluble Prmssian blue

Pmonred bw Ferric ion on ferroeyenic acid. Oxidation of ferroferroeyanic acid Acid on soluble Prussian blue Ferric ion on ferrocyanide ferricyhnide Ferrous ion Oxidation of ferrous ferroeyanide Redudion of ferric ferricyanide

+

Old F a m u l o Old Now6 HFe+++[Fc(CN)~]= Hydrogen ferrie ferro cyanide

Ncu, Formula Ht[[Fe(CN)#IZFel-

Berlinic* a d d

KFe+++[Fe(CN)81or KFet*[Fe(CNlal'

Potesriumferrieferracyanide or Potassium ferrour ferricyanide Ferric ferrocyanide

K+[[Pe(CN)cI~Fel~

Potaodumherlinate*

Fmouo ferrieyanide

K F e + + [ [ F e ( C N l a l ~ T e l ~Pota~sium ferrous or berlinrte* KFerti[[Fe(CN1~lhFelr

Ex-s ferric ion on ferroeyanide Fa+it[Fe(CNl+~ Ferric ion on soluble Frursian blue Excess ferrous ion on faricyanide F e i t + l F e ( C N ) ~ l F Ferrous ion on soluble Prussian blue * Balinnte = a mmpler of ferric ion end f e w i d e ion.

Insoluble Prussian blue Turnbull'r bluc

Nca Nnma

Fet*+[[Fe(CN)al~Fe]~F m i e berlinate*

~

~

although a polymeric formula is also possible. These complex ferrocyanides are produced by the theoretical quantity or excess of ferrocyanide. The so-called normal (or simple) ferrocyanides produced by excess of the metallic ion, such as Ag4[Fe(CN)s] should be viewed as heavy metal salts of the complex metallic ferrocyanides; in this case, silver triargentoferrocyanide, Ag+[AgJ[Fe(CN)s]]-.

sented by its structural unit containing one ferric ion and one ferrocyanide ion, which complex, for convenience, is termed berlinate ion. 4. Insoluble Prnssian blue is femc berlinate, Turnbull's blue, ferrous berlinate, or a double potassium ferrous berlinate. 5. The composition and color of other heavy metal ferrocyanides may be accounted for by assuming polynuclear complexes based on structural units deSUMMARY rived by considering the coordination numbers of the 1. Soluble Prussiau blue is KFe+++[Fe(CN)6]= metals involved. rather than KFe++[Fe(CN)$. 2. The color and composition of soluble Prnssiau blue are accounted for by assuming polynuclear The author wishes to express his indebtedness to complex formation between ferric ion and the nitrogen Professor Louis P. Hamrnett of Columbia University end of the cyanide groups present in ferrocyanide. for helpful discussions of the theoretical matter in 3. This supercomplex is highly polymerized owing this paper. to stereochemical restrictions, but may be repre-